>
> M> The latest version in CVS has some problems - it also seg
> M> faults under CBR for me and fails even the 6 frame testcase.wav
> M> test. Can people hold off commiting any more changes for a day
> M> or two until this is straightened out?
>
> Maybe that is my faults (because this two days, CVS changes are mainly
> done by me), but on my linux box, it is not happen... ? and "make test"
> has no problem too.
>
I found one of the floating point exceptions under CBR: the partition
band calculations in psymodel.c, before the spreading function is
applied, have to use npart_l_orig and npart_s_orig for the loop
bounds. (instead of npart_l and npart_s, which are the bounds after
the spreading function has been applied).
But I still get floating point exceptions in calc_xmin(). It
is the kind of problem that I can't track down in the debugger,
and by adding printf() functions it goes away, so I think it
is a strange memory problem.
Mark
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
